Incremental Computation of Billing Percentile Values in a Cloud Based Application Acceleration as a Service Environment

1. A method comprising: sampling a usage data of a network entity of an application acceleration as a service provider in intervals of five minutes using a processor and a memory, wherein each five minute interval constitutes a billing unit; a storage system configured to sort each billing unit of a customer based on bandwidth usage to determine a top five percent (5%) of samples in a period of a billing cycle; a process algorithm configured to designate the top 5% of samples in the period of the billing cycle as a burst bandwidth data; a processing unit configured to automatically calculate a 95th percentile value based on a next value in the billing cycle after the top 5% of samples in the billing cycle; a computation unit configured to incrementally compute the 95th percentile of each of a plurality of billable units for each of billing measurements for a large scale data associated with the network entity by computing the 95th percentile upon a newest set of data arrived to the network entity in each five minute interval; and a processing unit configured to determine a billing amount based on an incremental computation of the 95th percentile of each of the plurality of billable units for each of billing measurements for the large scale data associated with the network entity by a computation applied only upon the newest set of data arrived to the network entity in each five minute interval, wherein the computation does not require traversing through all the data for the billing cycle associated with each of the plurality of billable units.

2. The method of claim 1 : wherein the incremental computation includes processing only the newest set of data arrived at every run of the method, and wherein the newest set of data is a data that arrived from an Internet network between a current run and a previous run of a process.

3. The method of claim 2 further comprising: executing a sequence of processes to run a compute of all billing details across each of the plurality of billable units; persisting the sequence of processes into the storage system; and serving any request to view 95th percentile data from the storage system in real time whenever it is requested by the customer on an ad hoc basis; and generating a billing statement in real time whenever it is requested by the customer on the ad hoc basis through an incremental computation method.

4. The method of claim 3 further comprising: collecting a statistical data for all the network entities served; and reconciling a calculated amount across each of the plurality of billable units with an agreement with the customer at a commencement of an engagement with the service provider.

5. The method of claim 4 further comprising: applying the process algorithm to communicate the data into a distributed file system; and utilizing a Hadoop framework to provide fault tolerance to the data and to provide parallel computation of a billing network across all entities.

6. The method of claim 5 further comprising: defining a sequence of operations from all service models offered through the application acceleration as the service provider, wherein at least one sequence includes a Directed Acyclic Graph of map reduce applications; and chaining a map reduce functions into a SQL-like declarative data flow language, wherein the SQL-like declarative data flow language is Apache Pig.

7. The method of claim 6 further comprising: utilizing a persistence model comprising of bounded min max priority queue to attain the fault tolerance and replication; indexing a queue for each and every entity; resetting the queue for each network entity; and evoking a second level process to define a sequence of a map-reduce job that further processes the data.

8. The method of claim 7 : wherein the incremental computation to provide a very efficient and scalable method to automate a billing computation associated with the customer, and wherein the storage system to store results of incremental computations a-priori at short and defined intervals.

9. The method of claim 1 , wherein on every interval execution, a second level process: to analyze an input samples stored in a distributed file system, to aggregate a one minute data written by network nodes to 5 min data per each network entity, to search an indexed persistence model to fetch a min max priority queue for that entity and ingest bandwidth information from a network file data into a min max queue while preserving a structure of the min max priority queue, to initialize a bounded min max priority queue for each of the entities to zero at a first cycle of a billing period, to bound the min max priority queue with size equivalent to a 5% of a set of possible samples in the billing period plus one, to add a new set of values to the bounded min max priority queue, to reject any new value from the bounded min max priority queue if the new value is lesser than a smallest value in the bounded min max priority queue while the bounded min max priority queue is full, to evict the smallest value from the bounded min max priority queue if the new value is larger than the smallest value in the bounded min max priority queue while the bounded min max priority queue is full, to extract a min value of the bounded min max priority queue as the 95 th percentile value of the billing period, and to extract a max value of the bounded min max priority queue as a peak value of the billing period.

10. The method of claim 1 : wherein the usage data is sampled across each geographic location and each network entity associated with the customer of the application acceleration as the service provider.

11. The method of claim 10 : wherein the usage data is at least one of a calculation of bytes of data transferred from the network entity, bytes of data received by the network entity, a bits-per-second traffic rate sent from the network entity to a wide area network (WAN), and a bits per second traffic rate received to the network entity from a WAN.

12. The method of claim 11 : wherein the network entity is at least one of a point of presence (POP) node in the WAN, an aggregate network utilization estimation of the customer taken as a whole, and a logical network module deployed for the customer across in furtherance of the application acceleration as the service provider.

13. A method comprising: sampling a usage data of a network entity of an application acceleration as a service provider in intervals of five minutes, wherein each five minute interval constitutes a billing unit; a storage system configured to sort each billing unit of a customer based on bandwidth usage to determine a top nth percent of samples in a period of a billing cycle; a process algorithm configured to designate a top nth of samples in the period of the billing cycle as a burst bandwidth data; a processing unit configured to automatically calculate a 1-nth percentile value based on a next value in the billing cycle after the top nth percentile of samples in the billing cycle using a processor and a memory; a computation unit configured to incrementally compute the 1-nth percentile of each of a plurality of billable units for each of billing measurements for a large scale data associated with the network entity by computing the 1-nth percentile upon a newest set of data arrived to the network entity in each five minute interval using the processor and the memory; and a processing unit configured to determine a billing amount based on an incremental computation of the 1-nth percentile of each of the plurality of billable units for each of billing measurements for the large scale data associated with the network entity by a computation applied only upon the newest set of data arrived to the network entity in each five minute interval, wherein the computation does not require traversing through all the data for the billing cycle associated with each of the plurality of billable units, wherein the incremental computation includes processing only the newest set of data arrived at every run of the method, and wherein the newest set of data is a data that arrived from an Internet network between a current run and a previous run of a process.

14. The method of claim 13 further comprising: executing a sequence of processes to run a compute of all billing details across each of the plurality of billable units; persisting the sequence of processes into the storage system; and serving any request to view 1-nth percentile data from the storage system in real time whenever it is requested by the customer on an ad hoc basis; and generating a billing statement in real time whenever it is requested by the customer on the ad hoc basis through an incremental computation method.

15. The method of claim 14 further comprising: collecting a statistical data for all the network entities served; and reconciling a calculated amount across each of the plurality of billable units with an agreement with the customer at a commencement of an engagement with the service provider.

16. The method of claim 15 further comprising: applying the process algorithm to communicate the data into a distributed file system; and utilizing a Hadoop framework to provide fault tolerance to the data and to provide parallel computation of a billing network across all entities.

17. A method comprising: sampling a usage data of a network entity of an application acceleration as a service provider in intervals of five minutes using a processor and a memory, wherein each five minute interval constitutes a billing unit; a storage system configured to sort each billing unit of a customer based on bandwidth usage to determine a top nth of samples in a period of a billing cycle; a process algorithm configured to designate the top nth of samples in the period of the billing cycle as a burst bandwidth data; a processing unit configured to automatically calculate a 1-nth percentile value based on a next value in the billing cycle after the top nth of samples in the billing cycle; a computation unit configured to incrementally compute the 1-nth percentile of each of a plurality of billable units for each of billing measurements for a large scale data associated with the network entity by computing the 1-nth percentile upon a newest set of data arrived to the network entity in each five minute interval; a processing unit configured to determine a billing amount based on an incremental computation of the 1-nth percentile of each of the plurality of billable units for each of billing measurements for the large scale data associated with the network entity by a computation applied only upon the newest set of data arrived to the network entity in each five minute interval, wherein the computation does not require traversing through the usage data for the billing cycle associated with each of the plurality of billable units; applying the process algorithm to communicate the usage data into a distributed file system; and utilizing a Hadoop framework to provide fault tolerance to the usage data and to provide parallel computation of a billing network across all entities.

18. The method of claim 17 : wherein the incremental computation includes processing only the newest set of data arrived at every run of the method, and wherein the newest set of data is data that arrived from an Internet network between a current run and a previous run of a process.

19. The method of claim 18 further comprising: executing a sequence of processes to run a compute of all billing details across each of the plurality of billable units; persisting the sequence of processes into the storage system; and serving any request to view 1-nth percentile data from the storage system in real time whenever it is requested by the customer on an ad hoc basis; and generating a billing statement in real time whenever it is requested by the customer on the ad hoc basis through an incremental computation method.

20. The method of claim 19 further comprising: collecting a statistical data for all the network entities served; and reconciling a calculated amount across each of the plurality of billable units with an agreement with the customer at a commencement of an engagement with the service provider.